Project summary: Although altered cell metabolism is becoming an established hallmark of cancer, there remains a crucial need of new tools for quantitation of metabolites at single living cell level. In particular, due to lack of specific labels for metabolites, there is an unmet need for high-resolution imaging tools capable of mapping metabolites and small molecules (fatty acids, carbohydrates, amino acids) that play essential roles in pathogenesis of cancer. Supported by a R21 grant through the IMAT program, our team partially addressed this need via developing a multiplex stimulated Raman scattering (SRS) microscope, which enabled vibrational imaging of metabolites in live tumor cells and intact biopsies at the speed of 5 microseconds per spectrum. This R33 application aims to push the hyperspectral stimulated Raman imaging technology to the next level through (i) technical simplification and validation, (ii) developing a robust hyperspectral image segmentation framework, and (iii) integrating the SRS modality with a commercial spontaneous Raman microscope towards broad use by non-experts. We have assembled an inter-disciplinary team for the proposed development. The three specific aims are: (1) Developing an easy-to-operate, highly sensitive line-by-line hyperspectral SRS microscope and validate its capacity for cancer metabolic imaging at single cell level. (2) Establishing a feature analysis framework for segmentation of hyperspectral SRS images using the non-parametric Bayesian model. (3) Integrating and validating the stimulated Raman imaging modality on a spontaneous Raman microscope. By completing the proposed development and validation activities, we will have generated a highly novel spectroscopic imaging system broadly applicable to analysis of chemical contents of cells and tissues for discovery-driven research and marker- based precision diagnosis.